When a task fails and is retried, then fails again, the failure column in the logbook is overwritten by whatever the last failure is. It would be very beneficial to keep track of each failure individually so we can trace errors better on failed flows.